A “variator” is a device for transmitting rotary drive at a continuously variable ratio. Variators are used for example in continuously and infinitely variable transmissions for vehicles.
In a rolling traction type variator, drive is transmitted between rotating parts by virtue of traction at a rolling contact between them, and movement of one or more of the rotating parts provides for change in the variator drive ratio. The variator typically includes a first rotating part—a race—upon which runs a second rolling part—a roller.
The rollers and races can take various forms. Some rolling traction variators have conical races, movable along their axial direction to change ratio, with a wheel between them to transmit drive from one to the other. Another example uses spherical rollers rotatably mounted upon respective shafts. An example of this type is provided in U.S. patent application Ser. No. 10/141,652, published under US2002/0170799 and filed in the name of Donald C Miller.
A well known variator is the toroidal race type which comprises at least one pair of semi-toroidally recessed variator races mounted for rotation about a common axis and defining between themselves a generally toroidal cavity. The construction and operation of such variators is described in patents and patent applications held by Torotrak (Development) Ltd and others, including for example Torotrak's International Patent Application PCT/EP2006/)50860, published under no. WO2006/084905, the content of which is incorporated by reference.
In most cases rolling parts must be biased toward one another to provide traction. An important quantity in this regard is the traction coefficient μ, defined in the conventional manner as traction force divided by contact normal force, where the traction force is the force transmitted at the rolling contact and the contact normal force is the force with which the rolling parts are biased together, measured normal to the contact surfaces.
The rollers and races may be in direct mechanical contact, with drive being transmitted from one to the other through friction at the contact. Other rollers and races may be separated by a thin film of fluid (“variator fluid”). The variator fluid is typically jetted onto the rolling parts, and thus drawn into the region between them.
In a toroidal type variator, the contact normal force is typically provided by biasing one of the variator races toward the other. The force applied to the variator race (which determines the contact normal force but is not equal to it, since it is shared over multiple variator rollers and is not in general precisely parallel to the contact normal) is referred to as the “end load”. Some sophisticated variators use a hydraulic actuator to provide an end load which varies with the torque being handled by the variator. Some simple variators use a spring to provide a substantially constant end load. PCT/EP2006/050860, referred to above, provides an example of the latter.
There is a limit to the traction coefficient that can be sustained. If the required traction force becomes excessive in relation to the contact normal force, the result is an unacceptable degree of slippage at the rolling contact, which can result in damage to the variator. The limiting coefficient of traction—at which slippage becomes unacceptable—may depend upon several factors including, for example, the nature of the surfaces of the rolling parts and the elastohydrodynamic properties of the variator fluid, where present. A high value of the limiting coefficient of traction is desirable because it allows for a reduction in the end load. High end loads can reduce the variator's efficiency and reduce the effective life of the component parts especially the races and rollers.
The variator's rollers and races can cyclically suffer high Hertzian contact pressure. Also significant heat can be dissipated, creating potentially high temperatures. There may also be large tangential shear forces at their surface. These factors can lead to failure of the rollers and races, as explained in a paper entitled “Developing the Durability of a Dual-Cavity Full-Toroidal IVT Variator” (Adrian Lee, Jonathan Paul Newall: Torotrak (Development) Ltd, Yoshihiro Ono, Teruo Hoshino Koyo Seiko Co Ltd, SAE 2002 World Congress & Exhibition, March 2002, Detroit: Session: “Transmission and Driveline Systems Symposium (Part A)—IVT/CVT; Document Number 2002-01-0587, Book Number SP-1655). (Referred to as “Durability Paper”).
The Durability Paper describes a prior study of the factors affecting the fatigue life of the variator rollers and races. The rolling parts tested were wrought bearing steel with surfaces that were either ground or lapped. The paper explains that some of these parts underwent rolling contact fatigue, exhibited in two failure modes:
1. surface distress—“failure of rolling elements by the formation of glazed areas, followed by asperity scale microcracks which lead to asperity scale micro-spall craters”; and
2. spalling—“failure by the formation of macroscopic craters in the contact surface as a result of fatigue crack propagation in the Hertzian stress field” (the words in quotation marks are taken directly from the paper).
Both rendered the components unfit. The irregular running surfaces caused by surface distress were observed to lead to unacceptable vibration in the variator. In the trials described in the paper, testing was terminated at that point, but other trials have demonstrated that a bearing steel component which continues to be used after the onset of surface distress will suffer spalling.
Where surface distress did not occur (i.e., where the components functioned correctly, without failure) wear rate of the rolling surfaces of the rollers and races was so low as to make its assessment by the conventional method—measurement of weight loss—impractical. This of course is the result of the separation of the surfaces by the fluid film. A related observation (not detailed in the paper but demonstrated in other trials) is that variator rollers and races which have completed their design lifetime and even been tested to the point of destruction have often undergone so little wear as to retain upon their running surfaces the slight machining marks left by turning or grinding during their manufacture.
The study involved components with relatively rough running surfaces (0.13<Ra<0.23) and smoother components (Ra<0.1). The surface distress failure mode was observed only in the relatively rough components, whose lifetime was also shorter than the smoother parts. The Durability Paper implies that running surfaces need to be sufficiently smooth to prevent surface distress.
Variator's rolling parts are typically made as smooth as commercially possible to resist surface distress. The Durability Paper contained proposed alloys and surface treatments intended to provide compressive residual stress at the roller surface, to resist surface initiated cracking.
FIG. 6 to this application (not taken from the Durability Paper) is a micrograph showing the running surface of a wrought steel variator component made with roughness 0.13<Ra<0.23 that failed after being run for 74 hours at 1.8 GPa. Region 1000 shows the effects of surface distress. The surface distress has led to surface initiated spalling, forming a crater 1002. Bands 1004 and 1006 on either side of the running track are unaffected.
On the other hand, FIG. 7 to this application is a micrograph showing the running surface of a component having a surface roughness Ra<0.1, which ran for more than 2200 hours at the same pressure: 1.8 GPa. It can be seen that the region 1008, forming the centre of the running track, still shows the original machining features also seen in bands 1010, 1012 to either side of the track. Imperfection 1013 is the imprint of some debris. In peripheral running track regions 1014, 1016, in which film thickness was lowest, the material has been polished, its asperities having been being plastically deformed or smoothed in use.
Running surface properties, such as roughness, have a bearing on variator performance including the peak or limiting traction coefficient. While factors, such as choice of variator fluid can influence the limiting traction coefficient, it is believed that a suitable degree of roughness can assist in providing traction. However, increased roughness when using the conventional wrought steel parts appears to lead to premature fatigue failure.
International patent application PCT/NL00/00418, published under no. WO 00/77268 in the name of SKF Engineering and Research Centre B.V (“SKF”) discloses high alloy steel variator discs and/or rollers formed using then existing powder metallurgy manufacturing methods.
SKF states that “any method known in the art” may be used and identifies hot isostatic pressing followed by hot forging as a method of making the powdered metal. Hot isostatic pressing was a well known technique for forming high quality steel, involving the application of heat and high pressures to metal particles in a closed capsule, to create a very dense ingot without many internal voids.
The SKF surfaces are to be treated to achieve “a very high hardness”, by austenising and then quenching. The resulting component is said to have a very high surface hardness. For instance, the surface hardness was measured at 67 HRc. Somewhat similar content is found in SKF's International patent application PCT/NL00/00417, published under WO 00/79151.